IN THE UNITED STATES DISTRICT COURT FOR THE WESTERN DISTRICT OF WISCONSIN
TEAM SCHIERL COMPANIES and HEARTLAND FARMS, INC., on behalf of themselves and all others similarly situated,
OPINION and ORDER Plaintiffs,
v. 22-cv-580-jdp
ASPIRUS, INC. and ASPIRUS NETWORK, INC.,
Defendants.
This is a proposed class action about alleged antitrust violations in the healthcare industry. Plaintiffs Team Schierl Companies and Heartland Farms are businesses who offer self-insured health plans to their employees and purchase healthcare services from defendants Aspirus, Inc. and Aspirus Network, Inc. The court will refer to Aspirus, Inc. as “Aspirus” and Aspirus Network, Inc. as “ANI.” Aspirus is the dominant healthcare provider in north-central Wisconsin. ANI is a membership organization that operates a clinically integrated network (CIN) consisting of Aspirus-owned providers and contracted independent providers. ANI-CIN members delegate to ANI the authority to negotiate prices with payers on their behalf, and all ANI-CIN members agree to accept the same price from each payer as all other members. ANI also requires payers who wish to include any ANI-CIN member within their network to include all ANI-CIN members. Defendants say that this business model promotes care coordination and healthcare quality. Plaintiffs say that it’s an illegal price-fixing scheme in violation of § 1 of the Sherman Act. Two motions are before the court. First, defendants have moved to exclude a damages model produced by plaintiffs’ class certification expert Jeffrey Leitzinger, which purports to isolate the overcharges that class members paid as a result of defendants’ anti-competitive conduct. The court concludes that the methods underlying Leitzinger’s model are unreliable, so the motion to exclude will be granted. Second, Plaintiffs have moved to certify a damages class under Federal Rule of Civil Procedure 23(b)(3). Dkt. 185. That motion will be denied
because without Leitzinger’s damages model, plaintiffs have no way to measure damages on a class-wide basis.
BACKGROUND This case is about the payment of outpatient healthcare services.1 Healthcare services are generally paid for not by patients, but by third-party payers, such as commercial insurers or self-funded health plans. Commercial insurers provide health plans to individuals and entities in exchange for premium payments. Self-funded plans are offered by businesses, governments, or unions to their employees. Self-funded plans may contract with third-party
administrators (TPAs) to process claims and make payments, but they collect their own premiums and pay providers from their own funds. Plaintiffs Team Schierl Companies and Heartland Farms, Inc. offer self-funded health plans to their employees. They seek to represent a class of self-funded plans and commercial insurers who purchased outpatient healthcare services on behalf of patients from Aspirus and other providers who are members of the ANI-CIN.
1 The complaint alleged antitrust violations related to both outpatient and inpatient services, but after discovery, plaintiffs have narrowed their case to outpatient services only. Dkt. 186, at 12 n.2. The price that putative class members paid to Aspirus and other ANI-CIN members for outpatient services was set through negotiations between ANI and entities that the parties refer to as “network vendors.” Network vendors contract with healthcare providers to assemble provider networks. Many commercial insurers serve as their own network vendors, whereas
self-funded plans often use third-party network vendors to negotiate and assemble provider networks on their behalf. Network vendors represent large numbers of prospective patients, which they can use as leverage to negotiate a discounted “in-network” price from providers. ANI negotiates with a number of network vendors, but a majority of its contracts are negotiated with just five of them: The Alliance, Anthem Blue Cross Blue Shield, Security Health Plan, UnitedHealthcare, and the Aspirus HealthPlan. In this case, plaintiffs challenge two policies that ANI enforces in its negotiations with network vendors. The first policy is joint contracting: if a network vendor wants to include one
ANI-CIN member in a provider network, then it must include all ANI-CIN members, and the contract sets a single fee schedule for all ANI-CIN members. Second is exclusivity: ANI-CIN members are prohibited from negotiating their own contracts with a payer if ANI already has a contract with that payer. Defendants say that these policies are necessary to build a coordinated clinical network and deliver high quality healthcare. Plaintiffs say that they amount to illegal price-fixing, inflating the prices that plaintiffs and other payers pay for outpatient healthcare services.
ANALYSIS
Two motions are before the court. Defendants move to exclude several opinions from plaintiffs’ class certification expert Jeffrey Leitzinger, who used statistical techniques to estimate the overcharges the putative class members paid as a result of the challenged conduct. Dkt. 195. Plaintiffs move for class certification. Dkt. 185. The motion for class certification relies heavily on Leitzinger’s opinions, so the court will begin with the motion to exclude. See Am. Honda Motor Co. v. Allen, 600 F.3d 813, 816 (7th Cir. 2010) (“The court
must . . . resolve any challenge to the reliability of information provided by an expert if that information is relevant to establishing any of the Rule 23 requirements for class certification.”). A. Motion to exclude opinions of Jeffrey Leitzinger Jeffrey Leitzinger is an economist whose work focuses on antitrust economics. Leitzinger proposed a damages model to estimate the antitrust impact of the challenged conduct on the putative class members. Under Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 592–93 (1993) and Kumho Tire Company, Ltd. v. Carmichael, 526 U.S. 137, 147 (1999), the
court must ensure that proffered expert testimony meets the requirements of Federal Rule of Evidence 702. The gatekeeping function under Daubert essentially involves a three-part test: (1) the proffered expert must be qualified; (2) the expert’s methodology must be reliable; and (3) the expert’s testimony must be relevant. Gopalratnam v. Hewlett-Packard Co., 877 F.3d 771, 778 (7th Cir. 2017). Neither side disputes that Leitzinger’s opinions are relevant to both class certification and the merits of plaintiffs’ case. As for qualifications, the question is whether the expert has the necessary education and training to draw the conclusions he or she offers in the case at hand. See Hall v. Flannery, 840
F.3d 922, 926 (7th Cir. 2016). Defendants point out several times in their brief that Leitzinger is not a healthcare economist. Dkt. 196, at 10, 17, 23. But they don’t actually contend that Leitzinger has to be a healthcare economist to draw the conclusions he offers in his report. Nor does the court find that specific expertise in healthcare is necessary; Leitzinger’s conclusions rely on general economic methodologies, not anything specific to the healthcare industry. That leaves reliability. The test for reliability is necessarily flexible. Daubert identifies factors the court may consider when determining whether an expert’s testimony is
reliable—whether the expert’s technique has been tested, subjected to peer review and publication, analyzed for errors, or is generally accepted—but these factors “neither necessarily nor exclusively appl[y] to all experts or in every case.” Gopalratnam 877 F.3d at 779–80. The reliability inquiry focuses on the expert’s methodology; that is, whether the expert exercised “soundness and care” in reaching his opinions. Timm v. Goodyear Dunlop Tires N. Am., Ltd., 932 F.3d 986, 993 (7th Cir. 2019). The inquiry does not ask whether the expert’s ultimate conclusions are correct. “The soundness of the factual underpinnings of the expert’s analysis and the correctness of the expert’s conclusions based on that analysis are factual matters to be
determined by the trier of fact.” Gopalratnam, 877 F.3d at 781 (quoting Smith v. Ford Motor Co., 215 F.3d 713 (7th Cir. 2000)). Defendants contend that three of Leitzinger’s opinions should be excluded as unreliable: (1) his “yardstick” damages model, which estimated the average overcharge attributable to the challenged conduct; (2) his in-sample prediction analysis, which estimated the class-wide effects of the challenged conduct; and (3) his extrapolation analysis, which estimated damages for class members whose claims data was not produced in discovery. The court concludes that Leitzinger’s yardstick model is unreliable because Leitzinger failed to
adequately explain how the model isolated the effect of the challenged conduct on prices. The in-sample prediction analysis and extrapolation analysis both rely on the yardstick model, so the court will exclude them as well without considering defendants’ additional arguments. A yardstick model is a common methodology to estimate the impact of alleged anti-competitive conduct on prices. See, e.g., Blue Cross & Blue Shield United of Wisconsin v. Marshfield Clinic, 152 F.3d 588, 592–93 (7th Cir. 1998). A yardstick methodology uses prices in a market unaffected by the challenged conduct “as a yardstick against which outcomes in
the affected market can be compared.” Dkt. 191 (Leitzinger report), ¶ 31. After controlling for other, non-conspiratorial factors that might affect price, the difference in prices between the affected market and the yardstick is assumed to be the result of the challenged conduct. Id. For his analysis, Leitzinger selected a yardstick comprised of outpatient providers located in Wisconsin but outside north-central Wisconsin.2 Leitzinger explained that he excluded providers inside north-central Wisconsin because their prices may have been inflated by “umbrella effects” of ANI-CIN’s pricing. Dkt. 191, ¶ 34 n.50. He limited the yardstick to Wisconsin providers to “control for any pricing considerations that may be state specific.”
Dkt. 191, ¶ 34. To construct his yardstick model, Leitzinger relied on 58 million claims records from 2017–2024, which were produced by the five largest network vendors that contract with ANI: Anthem Blue Cross Blue Shield, United HealthCare, UMR (a subsidiary of United HealthCare that offers third-party administration for self-funded health plans), The Alliance, and Security Health Plan. Dkt. 191, ¶ 29; Dkt. 193, ¶ 2.3 Leitzinger compared ANI-CIN
2 Leitzinger used the definition of north-central Wisconsin proposed by plaintiff’s healthcare industry expert David Dranove. Dkt. 198 (Dranove report), ¶ 39. That definition includes the following counties: Iron, Vilas, Forest, Florence, Price, Oneida, Langlade, Lincoln, Taylor, Clark, Marathon, Shawano, Wood, Portage, Juneau, and Adams. 3 Leitzinger’s original expert report dated March 26, 2025, relied on 52 million claims records produced by the network vendors. Dkt. 191. On June 11, 2025, Leitzinger supplemented his report, updating his calculations to include 6 million additional claims records newly produced providers to yardstick providers by constructing a multiple linear regression model. The explanatory variable was ANI-CIN status, and the response variable was the price charged for outpatient services. Id. ¶¶ 30–35. Control variables included the type of outpatient service provided, the date of service, the place of service (such as a hospital, emergency room, or
doctor’s office); urbanicity of the service provider’s location; the provider’s physician specialty; the provider’s size; whether the provider was part of a hospital system, standalone hospital, or standalone private practice; and characteristics of the insurance plan and network vendor. Id. ¶¶ 32–33. After controlling for these variables, Leitzinger’s model estimated an average overcharge of 18.9 percent for ANI-CIN providers compared with yardstick providers. Id. ¶ 35. Leitzinger concluded that this 18.9 percent overcharge was attributable to the challenged conduct in this case. Defendants concede that a yardstick model is a generally accepted method of computing
antitrust damages. Dkt. 196, at 20; see also Marshfield Clinic, 152 F.3d at 592–93; Fishman v. Est. of Wirtz, 807 F.2d 520, 551 (7th Cir. 1986); Conwood Co., L.P. v. U.S. Tobacco Co., 290 F.3d 768, 793 (6th Cir. 2002). But that doesn’t mean that every yardstick model is necessarily admissible. The reliability of any specific yardstick model depends on whether it can reasonably estimate the prices that would have existed in the affected market if not for the anti-competitive conduct. City of Rockford v. Mallinckrodt ARD, Inc., No. 3:17-cv-50107, 2024 WL 1363544, at 7 (N.D. Ill. Mar. 29, 2024); Fishman, 807 F.2d at 550–51. The model must be consistent with plaintiffs’ theory of antitrust liability; in other words, it must measure damages attributable to
by Anthem Blue Cross Blue Shield. Dkt. 193. The updated analysis did not change any of Leitzinger’s conclusions. Leitzinger explains his conclusions in his original report, so the court will cite that report for Leitzinger’s reasoning, but use the calculations from the supplemental report. the specific conduct that the plaintiffs are challenging. Comcast Corp. v. Behrend, 569 U.S. 27, 35 (2013). And the expert must provide sound reasoning for both the yardstick he selects and the variables he includes as controls; otherwise, his conclusion that the model estimates the effects of the challenged conduct is mere ipse dixit. Mallinckrodt, 2024 WL 1363544, at *10.
Defendants contend that Leitzinger’s yardstick model is unreliable because Leitzinger didn’t explain why his yardstick is an appropriate control and because Leitzinger didn’t account for key variables known to affect the price of outpatient care. The court agrees with defendants on both counts. 1. Selection of appropriate yardstick As Leitzinger explained in his report, a yardstick model attempts to estimate price in a but-for world where everything is the same except for the alleged anti-competitive conduct. See
Dkt. 191, at 18. The yardstick stands in for the but-for world, so the extent to which the yardstick approximates the target market absent the challenged conduct is key to its reliability. Defendants contend that Leitzinger’s yardstick is unreliable because he failed to ascertain whether it was “tainted” by the challenged conduct, meaning that providers in the yardstick group had joint contracting or exclusivity policies like ANI-CIN’s. Plaintiffs raise two arguments in response. First, they say that Leitzinger did provide sound reasons for assuming that the yardstick was free of the challenged conduct. Second, they say that even if the yardstick was tainted, that wouldn’t make the model itself unreliable.
Leitzinger gave three reasons in his expert report why he assumed that the providers in the yardstick did not engage in the challenged conduct. Dkt. 191, ¶ 34. First, he pointed to the deposition of Andrea Lathers, corporate witness for Anthem Blue Cross Blue Shield of Wisconsin, who testified that it was “not common” in Wisconsin for clinically integrated networks to negotiate a single contract so that every provider in the network receives the same price for the same service. Dkt. 178 (Lathers Dep. 103:23–104:3). Second, Leitzinger noted that published FTC guidance during the class period “cautioned against” the use of exclusive outpatient networks representing a large number of outpatient providers. Third, Leitzinger
analyzed whether prices for outpatient services at hospitals owned by three health systems in the yardstick—Froedtert, Ascension, and Mayo Clinic—aligned with prices at independent outpatient providers in the same county as the hospitals. Leitzinger found that the independent providers’ prices were on average 23% lower than the hospitals’ prices, which he took as evidence that these hospital systems were not engaging in joint price setting in the same way that ANI-CIN did.4 Dkt. 191, ¶ 34 n.53. The court does not find these to be sound reasons for assuming that yardstick providers did not engage in the challenged conduct. As for the first reason, Leitzinger misrepresented
Lathers’ testimony in his report. Although Lathers initially testified that it was “not common” for clinically integrated networks to negotiate a single price for all providers, she reversed that testimony only moments later, explaining that “[o]ther clinically integrated networks in Wisconsin also negotiate in a similar manner [to ANI].” Dkt. 178 (Lathers Dep. 105:2–4). She then named five networks in Wisconsin that engage in this type of conduct: OakLeaf Surgical clinically integrated network, ThedaCare ACO, Bellin Physician Partners, Independent Physician Network, and Wisconsin IPA. Id. at 105:12–24. All of these providers are part of
4 Leitzinger did the same analysis for Aspirus hospitals and independent ANI-CIN members in the same county as the Aspirus hospitals. The price difference between the hospitals and the independent providers was close to zero, which is consistent with ANI-CIN’s joint price setting policy. Leitzinger’s yardstick. But Leitzinger failed to investigate any of these providers’ price-setting policies; instead, he inexplicably relied on Lather’s initial, apparently incorrect, representation. As for Leitzinger’s second reason, he does not explain why the FTC guidance suggests that the providers in the yardstick were not engaged in the challenged conduct. Leitzinger
appears to have assumed that the yardstick providers, unlike ANI, complied with the FTC guidance. But he provides no basis for that assumption. As for the third reason, Leitzinger’s analysis of the differences in prices between hospital providers at Froedtert, Ascension, and Mayo, and independent providers in the same counties might be evidence that these three health systems do not engage in the challenged conduct. But Leitzinger provides no reason why he conducted this analysis for only these three health systems, or why his conclusion that these three health systems don’t set joint prices can be extrapolated to the rest of the yardstick.
Leitzinger failed to provide a sound reason for assuming that his yardstick was free of the challenged conduct, so that assumption rested on shaky ground from the beginning. But even more problematically, Leitzinger failed to re-evaluate his assumption in the face of evidence from defendants’ expert Laurence Baker that at least one yardstick provider did in fact engage in the challenged conduct. In one test of Leitzinger’s yardstick model, Baker applied the model to data from ThedaCare, a healthcare system located in northeastern Wisconsin that was part of the yardstick. Dkt. 194 (Baker report), ¶¶ 127–30. The model estimated a 16.1% overcharge for ThedaCare. Adopting Leitzinger’s assumption that the yardstick providers did
not engage in the challenged conduct, Baker interpreted this as a sign of the model’s weakness, reasoning that ThedaCare should have shown no overcharge because it was part of the yardstick. Id. But in his rebuttal report, Leitzinger wrote that the model likely estimated an overcharge for ThedaCare because ThedaCare did in fact engage in the challenged conduct to at least some extent. Dkt. 192, ¶¶ 60–63. To support this assertion, Leitzinger pointed to materials produced in discovery that suggest that ThedaCare also sets a single set of prices for outpatient providers within its clinical network. Id. ¶ 61. (Although Leitzinger didn’t mention
it, ThedaCare was also one of the provider networks Lathers identified as setting joint prices in a similar manner to ANI.) Leitzinger’s conclusion about ThedaCare amounts to an admission that the yardstick is, to some degree, tainted by the challenged conduct. Leitzinger’s admission directly undermines his initial assumption that the yardstick providers did not engage in the challenged conduct. Nevertheless, Leitzinger failed to revisit his initial assumption that yardstick providers did not engage in the challenged conduct. At minimum, Leitzinger should have analyzed whether the four other providers that Lathers identified in her deposition engaged in joint price setting, but he didn’t do that, nor did not
explain how a tainted yardstick would likely affect the model as a whole. Leitzinger did exclude ThedaCare from the yardstick and re-run the model, noting that the estimated overcharge for ANI increased from 18.9 percent to approximately 20 percent. Dkt. 192, ¶ 65. But that fix doesn’t account for other providers in the yardstick who may also have engaged in the challenged conduct even though Leitzinger didn’t know it. That brings the court to plaintiffs’ second argument, which is that even if other providers in the yardstick engaged in the challenged conduct, the effect would be harmless because it would only render the model overly conservative. Plaintiffs are correct that courts
generally do not require experts to demonstrate that their yardstick is entirely free from the anti-competitive conduct. In re Keurig Green Mountain Single-Serve Coffee Antitrust Litig., No. 14-md-2542, 2025 WL 354671, at *11 (S.D.N.Y. Jan. 30, 2025) (“[I]t would create an unworkable and unfair standard to require . . . that the yardstick selected perfectly reflect the affected company but for the anticompetitive conduct.”) (collecting cases). When a tainted yardstick causes the model to underestimate antitrust impact rather than overestimate it, courts generally decline to exclude the model on that basis. E.g In re Packaged Seafood Prods.
Antitrust Litig., 332 F.R.D. 308, 327 (S.D. Cal. 2019); Fond Du Lac Bumper Exch., Inc. v. Jui Li Enter. Co., Ltd., No. 09-CV-0852, 2016 WL 3579953, at *9 (E.D. Wis. June 24, 2016). But that general principle isn’t helpful here, because it’s not clear that taint in this particular yardstick would render the model more conservative. Crucially, Leitzinger never actually says in any of his reports that taint in the yardstick would necessarily make his model more conservative. Leitzinger said that in his deposition, Dkt. 190 (Leitzinger Dep. 144:14–145:16), but experts must set forth all of their opinions in their written reports; they cannot cure a deficient report via deposition. Fed. R. Civ. P.
26(a)(2); Salgado by Salgado v. Gen. Motors Corp., 150 F.3d 735, 741 n.6 (7th Cir. 1998). Leitzinger did observe an increase in the model’s estimated overcharge from 18.9 percent to 20 percent when he excluded ThedaCare, which implies that ThedaCare’s inclusion did in fact make the model more conservative. Dkt. 192, ¶ 65. But he did not extrapolate that finding to any providers other than ThedaCare or express any general opinion about how taint in the yardstick would affect his model as a whole. Plaintiffs say that it’s simply a general rule that taint in an antitrust yardstick model makes the model overly conservative. That argument makes some conceptual sense:
anti-competitive conduct generally leads to higher prices, so a tainted yardstick would be expected to have higher prices than the “but-for” world it is intended to estimate. But the court can’t assume that the general rule applies in this case, because defendants have introduced evidence to the contrary. Baker, defendants’ healthcare economics expert, opined in his expert report that price effects of the anti-competitive conduct in this case aren’t necessarily one-directional; they could either increase or decrease prices. Specifically, Baker wrote that joint contracting can reduce “transaction costs” associated with negotiating healthcare
networks, which in turn could reduce prices. Dkt. 194 (Baker report), ¶¶ 36–44. In light of Baker’s opinion, and without any opinion of any kind on the issue from Leitzinger, the court cannot assume that a tainted yardstick in this case would be harmless. Leitzinger’s complete failure to analyze the taint issue is therefore tantamount to ignoring an obvious alternative explanation for his findings. 2. Control variables Defendants’ second critique of Leitzinger’s yardstick relates to the control variables that
Leitzinger selected. Control variables account for factors other than the anti-competitive conduct that might explain differences in price between the defendant and the yardstick group. See Marshfield Clinic, 152 F.3d at 593. If a model fails to properly control for alternative factors, then it cannot support a reasonable inference that any price differences between the defendant and the yardstick are caused by the anti-competitive conduct. Defendants argue that Leitzinger’s model fails to control for market share and quality, which the Seventh Circuit has recognized as two of “the most important factors” affecting the price of healthcare services. Dkt. 196, at 22–23 (citing Marshfield Clinic, 152 F.3d at 593). In
response, plaintiffs say that the selection of control variables typically goes to the probative weight of a regression analysis, not to admissibility. Dkt. 202, at 30 (citing Manpower, Inc. v. Ins. Co. of Pennsylvania, 732 F.3d 796, 808 (7th Cir. 2013)). Plaintiffs acknowledge that a regression model may be inadmissible if an expert entirely fails to control for salient factors other than the target conduct that might affect the dependent variable. E.g., Mallinckrodt, 2024 WL 1363544, at *8 (excluding an antitrust damages model in which the expert “made no effort to control for any other factors that might have affected” the challenged product’s price). But plaintiffs point out that Leitzinger controlled for numerous factors that might affect the price
of outpatient healthcare services, including factors that plaintiffs say might be indicative of market share or quality, such as provider size, specialty, and whether the provider was part of a hospital system, standalone hospital, or private practice. Dkt. 191, ¶¶ 32–33. Accordingly, plaintiffs argue that defendants’ critique amounts to a disagreement between the parties’ experts about whether the controls properly account for all relevant factors, which is an issue of weight, not admissibility. But there’s a bigger problem with Leitzinger’s analysis, which is that Leitzinger himself fails to consistently explain what he controlled for and consequently, what his model measures.
A fundamental assumption of Leitzinger’s model is that it isolates the effects of the anti-competitive conduct at issue in this case. But Leitzinger appears to take a position in his rebuttal report that is inconsistent with that assumption. In that rebuttal report, Leitzinger attempted to explain why two of the providers in his yardstick, ThedaCare and Mayo Clinic, showed significant overcharges when run through his model. Dkt. 192, ¶¶ 63–64. He suggested that ThedaCare may be engaging in the anti-competitive conduct, as discussed earlier, but he also opined that these two providers might show overcharges because they have a reputation for providing particularly high-quality care. Id. But that reasoning doesn’t make sense—if
Leitzinger’s model truly controlled for relevant, non-conspiratorial factors that might affect price, then an overcharge calculated by the model shouldn’t be explainable based on differences in the quality of care. Leitzinger’s apparent concession that his model doesn’t account for quality of care is a fatal problem, because it undermines his bottom-line conclusion that the overcharge he calculated for ANI-CIN is attributable to the challenged conduct. That’s because Leitzinger doesn’t explain why ThedaCare or Mayo Clinic’s overcharges could be due to quality, but
ANI-CIN’s overcharge is due to the challenged conduct. Based on Leitzinger’s own explanation of his model, it seems just as reasonable that the overcharge the model calculated for ANI-CIN could have been due to quality effects. Leitzinger’s failure to account for this alternative explanation makes his conclusion fundamentally unreliable. Plaintiffs raise two arguments in their response brief in support of Leitzinger’s model, neither of which are persuasive. First, they point out that Leitzinger’s model has a high “adjusted r-squared” of 89.7 percent, a statistic that measures how much of the variation in the dependent variable is explained by the independent variables in the model. Dkt. 202, at
17; see also Federal Judicial Center, Reference Manual on Scientific Evidence, Reference Guide on Statistics, 404 (3d ed. 2011). But a high r-squared value does not imply that the model accounts for relevant control variables such as quality of care, nor does Leitzinger say that it does. A model that omits a key variable can suffer from systematic bias and still show a strong fit between the independent and dependent variables. See James H. Stock and Mark W. Watson, Introduction to Econometrics 238 (3d ed. 2015), Dkt. 225-1. Second, plaintiffs say that another analysis Leitzinger conducted rules out the possibility that ANI-CIN’s high prices were due to differences in quality as opposed to the
challenged conduct. Leitzinger created a differences-in-differences (DiD) model, which measures the effect of a specific event on a target variable by comparing a treatment group to a control group before and after the event occurs. Dkt. 191 (Leitzinger report), ¶¶ 40–44. In this case, Leitzinger analyzed prices for a group of healthcare providers who formerly belonged to the Ascension health system both before and after that system was acquired by ANI-CIN in 2021. Leitzinger then compared those prices to prices for other Wisconsin providers during the same period of time, using similar control variables as his yardstick analysis. Leitzinger
observed that relative prices increased significantly among the Ascension providers almost immediately after they switched to ANI-CIN. Id. ¶ 44. Leitzinger said that the DiD analysis refutes any contention that high prices in ANI-CIN were due to increased quality of care, because there wouldn’t have been enough time for any significant increase in quality prior to the prices increasing. Dkt. 192 (Leitzinger rebuttal report), ¶¶ 82–83. The DiD analysis does not rescue Leitzinger’s yardstick model, for two reasons. First, as Leitzinger acknowledged in his reports, the DiD model is not a comprehensive damages model. It included only a small set of providers who all formerly belonged to the same health
system, which could introduce its own set of biases. The DiD model may provide some corroboration for Leitzinger’s other conclusions, but it doesn’t obviate the need for Leitzinger to include appropriate control variables in his damages model to account for price effects unrelated to the challenged conduct. Second, the mere fact that prices increased almost immediately after the Ascension providers joined ANI-CIN doesn’t foreclose the possibility that the price increases were due to quality effects. As Leitzinger noted elsewhere in his rebuttal report, quality-related price effects can reflect a health system’s reputation for quality. Id. ¶ 64 (explaining that Mayo Clinic is “nationally renowned for quality care and research” and has a
“high quality reputation.”). So even if the quality of the Ascension providers couldn’t have improved in such a short time after joining ANI-CIN, prices could have increased based solely on ANI-CIN’s reputation for quality care. The bottom line is that an anti-trust regression model is only reliable if it reasonably estimates the price plaintiffs would have paid in the absence of the anti-competitive conduct. To do so, the model must use a valid comparator group and adequately control for price effects caused by factors other than the challenged conduct. Mallinckrodt, 2024 WL 1363544, at *10.
Leitzinger failed to provide a consistent or well-reasoned explanation for why his model meets either of these criteria. His yardstick group is tainted by the same anti-competitive behavior charged against ANI, and his model doesn’t account for quality of care or market share. These are not merely debatable shortcomings that go to the weight of his evidence. His assertion that his model isolates and measures antitrust impact is inadmissible because it is not based on reliable methods. The court will grant defendants’ motion to exclude the yardstick model under Rule 702. The court will also exclude Leitzinger’s in-sample and extrapolation analyses because they both rely on the yardstick model.
B. Motion for class certification The requirements for class certification under Rule 23 are well established: (1) the scope of the class as to both its members and the asserted claims must be “defined clearly” using “objective criteria,” Mullins v. Direct Digital, LLC, 795 F.3d 654, 657 (7th Cir. 2015); (2) the class must be sufficiently numerous, include common questions of law or fact, and be adequately represented by plaintiffs (and counsel) who have claims typical of the class, Fed. R. Civ. P. 23(a); and (3) the class must meet the requirements of at least one of the types of class
actions listed in Rule 23(b). Plaintiffs ask for certification of a damages class under Rule 23(b)(3), which applies if “the questions of law or fact common to class members predominate over any questions affecting only individual members” and “a class action is superior to other available methods for fairly and efficiently adjudicating the controversy.” The ultimate question in a Rule 23(b)(3) class is whether “judicial economy from consolidation of separate claims outweighs any concern with possible inaccuracies from their being lumped together in a single proceeding for decision by a single judge or jury.” Mejdrech v. Met-Coil Sys. Corp., 319 F.3d 910, 911 (7th
Cir. 2003); see also Suchanek v. Sturm Foods, Inc., 764 F.3d 750, 761 (7th Cir. 2014) (“Ultimately, the court must decide whether classwide resolution would substantially advance the case.”). Plaintiffs seek certification of the following class under Rule 23: All Payors whose funds were used to pay Defendants and/or their Co-Conspirators for in-network outpatient professional services provided in North-Central Wisconsin, during the period October 11, 2018, up to and including June 30, 2023 (the “Class Period”).5
Excluded from this Class are (1) individuals or entities whose only payments to Defendants were co-pays, co-insurance, and/or other out-of- pocket payments for out-of-network claims, and (2) individuals or entities that paid for only one claim. Also excluded from this Class are Aspirus, ANI, Aspirus Health Plan, and their officers, directors, management, employes, subsidiaries, or affiliates, judicial officers and their personnel, and all federal governmental entities. Dkt. 186, at 12. Essentially, the class consists of commercial insurers and self-funded health plans who paid ANI-CIN members for outpatient services during the relevant period. Plaintiffs also propose an alternative class, which is the same as the above, but includes only entities “who used The Alliance, Anthem, Security Health Plan, United HealthCare, and/or United Healthcare Management Resources as a Network Vendor and/or [third-party administrator].”
5 Plaintiffs use the term “co-conspirators” in their briefs to refer to non-Aspirus members of the ANI-CIN. Dkt. 186, at 12 n.3. These entities provided claims data during discovery, so the calculation of antitrust damages for the alternative class doesn’t rely on Leitzinger’s extrapolation analysis. Defendants raise four issues in opposition to class certification. First, they contend that plaintiffs cannot satisfy Rule 23(b)(3)’s predominance requirement because they lack common
evidence of antitrust impact. Second, they contend that the proposed class definition includes entities who did not directly pay ANI-CIN members for healthcare. Third, they contend that a class action would not be superior in this case because there are individualized issues related to standing and damages. Fourth, they contend that named plaintiff Heartland Farms cannot adequately represent the class because its representative testified that it “does not care” about the interests of commercial insurers. The predominance issue is dispositive, so the court will deny the motion for class certification on that basis without reaching the other issues. The exclusion of Leitzinger’s damages model is fatal to plaintiffs’ ability to satisfy the
predominance requirement under Rule 23(b)(3). Predominance is satisfied when “common questions represent a significant aspect of a case and . . . can be resolved for all members of a class in a single adjudication.” Messner v. Northshore Univ. HealthSystem, 669 F.3d 802, 815 (7th Cir. 2012) (quoting 7AA Charles Alan Wright & Arthur R. Miller, Federal Practice & Procedure § 1778 (3d ed. 2011)). Plaintiffs are not required to show complete commonality of damages, but the rule does require plaintiffs to put forth a common methodology that has the ability to measure damages on a class-wide basis; otherwise “[q]uestions of individual damage calculations will inevitably overwhelm questions common to the class.” Comcast Corp., 569 U.S.
at 34; see also Butler v. Sears, Roebuck & Co., 727 F.3d 796, 799 (7th Cir. 2013) (“Comcast holds that a damages suit cannot be certified to proceed as a class action unless the damages sought are the result of the class-wide injury that the suit alleges.”). As plaintiffs concede in their reply brief, their only class-wide damages evidence is Leitzinger’s yardstick model, which the court has excluded. Dkt. 219, at 47. Plaintiffs point to other, qualitative evidence of antitrust impact arising from the challenged conduct, including internal ANI documents stating that the challenged conduct prevented “price wars,”
Dkt. 187-29, and market analyses conducted by health economist David Dranove, concluding that the challenged conduct reduced competition. Dkt. 198. But none of this evidence can measure class-wide damages as required by Comcast. The damages model is the heart of this case; without it, individual damage calculations will “inevitably overwhelm questions common to the class.” Comcast, 569 U.S. at 34; see also City of Rockford v. Mallinckrodt ARD, Inc., No. 3:17- CV-50107, 2024 WL 1363544, at *10 (N.D. Ill. Mar. 29, 2024) (“In this case, as in others where it is essential, “[n]o damages model, no predominance, no class certification.” (quoting In re Rail Freight Fuel Surcharge Antitrust Litig., 934 F.3d 619, 626 (D.C. Cir. 2019)). Plaintiff’s
motion for class certification will be denied.
NEXT STEPS The court’s ruling on the admissibility of Leitzinger’s opinions and on class certification has significant implications for summary judgment and the court recognizes that plaintiffs will likely pursue their right to interlocutory appeal. Accordingly, the court will stay the dispositive motions deadline. The parties will have two weeks to file a joint status report advising the court how they wish to proceed. ORDER IT IS ORDERED that: 1. Plaintiffs’ motion for class certification, Dkt. 185, is DENIED. 2. Defendants’ motion to exclude expert testimony of Jeffrey Leitzinger, Dkt. 195, is GRANTED. Leitzinger’s yardstick damages model, in-sample prediction analysis, and extrapolation analysis are EXCLUDED. 3. The dispositive motions deadline is STAYED. The parties have until January 5, 2026 to file a joint status report advising the court how they wish to proceed. Entered December 19, 2025. BY THE COURT:
/s/ ________________________________________ JAMES D. PETERSON District Judge